Automatic calibration method and system for precision balance based on multi-sensor fusion

Abstract

The application relates to the technical field of balance calibration, in particular to a precise balance automatic calibration method and system based on multi-sensor fusion, which comprises the following steps: continuously acquiring temperature data, mass monitoring data and inclination data of the balance; analyzing the temperature data deviation degree, fluctuation degree and rising trend in each weighing process, the random fluctuation degree of the inclination data and the irregular oscillation degree of the mass monitoring data in the attenuation process; combining the synchronous change degree between the temperature influence degree and the dynamic unbalance characteristics in each weighing process and the previous weighing processes; constructing comprehensive abnormal characteristic values of each weighing process; optimizing the proportional term parameters of the PID controller in the next weighing process; and automatically calibrating the weighing result of the next weighing process. The application improves the stability and accuracy of the weighing result by self-adaptive optimization of the proportional term parameters in the weighing process.

Description

Technical Field

[0001] This application relates to the field of balance calibration technology, specifically to an automatic calibration method and system for precision balances based on multi-sensor fusion. Background Technology

[0002] Electronic analytical balances are high-precision mass measuring instruments. They are accurate and fast in weighing, simple to operate, and have good stability. They also have multiple functions such as automatic calibration and fault self-diagnosis, and are therefore widely used in scientific research, medicine, chemical industry and other fields.

[0003] Electronic analytical balances typically employ the electromagnetic force feedback zero-point method for weighing. The electromagnetic force balance sensor and regulating circuit form a closed-loop control system, outputting stable mass weighing data only when the system is in equilibrium. The speed and stability of the closed-loop control system's balance adjustment are crucial for achieving rapid and accurate weighing. However, as a high-precision measuring instrument, the electronic analytical balance has complex internal components and circuitry, making it susceptible to insufficient measurement accuracy due to changes in ambient temperature and the heating of internal overcurrent components. Furthermore, prolonged operation can alter the performance of the permanent magnet and the parameters of other components, making it difficult to adapt to complex and changing working environments, ultimately leading to unstable or inaccurate weighing results. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide an automatic calibration method and system for precision balances based on multi-sensor fusion. The specific technical solution adopted is as follows: In a first aspect, embodiments of this application provide an automatic calibration method for a precision balance based on multi-sensor fusion, the method comprising the following steps: During the weighing process using the balance, the temperature data, mass monitoring data, and tilt angle data of the x-axis and y-axis are continuously acquired. By analyzing the differences in temperature data relative to the pre-obtained optimal working temperature, the degree of temperature fluctuation, and the characteristics of temperature rise during each weighing process, a first feature value is constructed to characterize the degree of influence of temperature on the weighing results during each weighing process. Based on the degree of random fluctuation of the tilt angle data of the x-axis and y-axis during each weighing process, and the degree of irregular oscillation of the quality monitoring data during attenuation during each weighing process, the second characteristic value of each weighing process is constructed. Combined with the degree of synchronous change between the first characteristic value and the second characteristic value, and the first characteristic value of each weighing process, the comprehensive abnormal characteristic value of each weighing process is determined. By utilizing comprehensive abnormal characteristic values, the proportional term parameter of the PID controller is optimized for the next weighing process, thereby automatically calibrating the weighing results for the next weighing process.

[0005] As a preferred embodiment, the method for constructing the first characteristic value of each weighing process is as follows: A temperature difference coefficient is constructed based on the difference between the temperature data during each weighing process and the pre-obtained optimal working temperature. Calculate the standard deviation of all temperature data during each weighing process; Based on the temperature rise trend characteristics of each weighing process and the preceding weighing processes, a trend rise factor is constructed. The first characteristic value of each weighing process is positively correlated with the temperature difference coefficient, the standard deviation, and the trend increase factor.

[0006] In a preferred embodiment, the temperature difference coefficient refers to the average of the absolute differences between all temperature data and the optimal working temperature during each weighing process and within the preset time period prior to it.

[0007] As a preferred embodiment, the method for constructing the trend-upward factor is as follows: The average temperature values ​​of each weighing process and the preceding several weighing processes are arranged in chronological order and recorded as the temperature average value sequence of each weighing process. The Mankendall test algorithm was used to perform trend analysis on the temperature mean series of each weighing process, and the output S statistic was recorded as the trend increase factor of each weighing process.

[0008] As a preferred embodiment, the method for constructing the second characteristic value of each weighing process is as follows: The permutation entropy of the first difference values ​​of the inclination angle data of the x-axis and y-axis during each weighing process is calculated separately. Based on the degree of irregular oscillation of mass monitoring data during decay in each weighing process, an irregular decay coefficient is constructed. The second characteristic value of each weighing process is positively correlated with the permutation entropy and the irregular decay coefficient.

[0009] As a preferred embodiment, the method for constructing the irregular attenuation coefficient is as follows: The time period in each weighing process in which the range of quality monitoring data within a continuous preset monitoring time is less than the preset fluctuation threshold for the first time is obtained, and the last moment of the obtained time period is recorded as the steady state moment. The average value of all quality monitoring data after the steady-state moment in each weighing process is recorded as the target mass. Curve fitting is performed on the mass monitoring data from the moment the target mass is first reached to the steady state moment during each weighing process, and then the extreme points in the obtained fitted curve are obtained. The coefficient of variation of the absolute difference between the amplitudes of all adjacent extreme points is denoted as the irregular attenuation coefficient of each weighing process.

[0010] As a preferred embodiment, the method for constructing the comprehensive abnormal characteristic values ​​of each weighing process is as follows: The first and second characteristic values ​​of each weighing process and the preceding several weighing processes are arranged in chronological order and denoted as the first characteristic sequence and the second characteristic sequence of each weighing process. Calculate the maximum mutual information coefficient between the first feature sequence and the second feature sequence; The comprehensive abnormal characteristic value of each weighing process is positively correlated with the maximum mutual information coefficient, the first characteristic value of each weighing process, and the second characteristic value of each weighing process.

[0011] In a preferred embodiment, the optimization formula for the proportional term parameter of the PID controller during the next weighing process is as follows: In the formula, For the (i+1)th weighing process, the optimized proportional term parameter of the PID controller is used. , These are the preset minimum and maximum values ​​of the proportional parameter of the PID controller, respectively. It is the normalized value of the comprehensive abnormal characteristic value of the i-th weighing process.

[0012] As a preferred embodiment, the specific process of automatically calibrating the weighing result of the next weighing process is as follows: In the next weighing process, the PID controller uses the calculated optimized proportional term parameters to perform balance adjustment, thereby achieving automatic calibration of the weighing results.

[0013] Secondly, embodiments of this application also provide an automatic calibration system for a precision balance based on multi-sensor fusion, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any of the above-described automatic calibration methods for a precision balance based on multi-sensor fusion.

[0014] This application has at least the following beneficial effects: This application addresses the problem of unstable and inaccurate weighing results in electronic analytical balances due to the susceptibility of environmental temperature changes and internal overcurrent components to heating. By deeply analyzing the comprehensive offset characteristics and trends of temperature data during the weighing process, a first characteristic value is constructed, which can assess the impact of temperature data changes on the weighing results in each weighing cycle. By analyzing the random fluctuation characteristics of tilt angle data and the irregular attenuation characteristics of mass monitoring data during each weighing cycle, a second characteristic value is constructed, which can assess the dynamic balance performance of the balance in each weighing cycle. Furthermore, by evaluating the synchronicity between abnormal temperature characteristics and abnormal dynamic balance characteristics, a comprehensive abnormality characteristic value is constructed, which can assess the comprehensive abnormality of the weighing results under multiple disturbances. This value is then used to optimize the proportional term parameters of the PID controller in the next weighing cycle, enabling the electronic analytical balance to adapt to complex and changing working environments and real-time changes in the performance of internal components during weighing, achieving automatic calibration of the weighing results, and improving the stability and accuracy of the weighing results. Attached Figure Description

[0015] Figure 1 A flowchart illustrating the steps of an automatic calibration method for a precision balance based on multi-sensor fusion, provided in one embodiment of this application; Figure 2 A flowchart illustrating the acquisition of comprehensive abnormal feature values ​​for each weighing process, provided in one embodiment of this application. Detailed Implementation

[0016] Please see Figure 1 The diagram illustrates a flowchart of an automatic calibration method for a precision balance based on multi-sensor fusion, according to an embodiment of this application. The method includes the following steps: Step 1: During the weighing process using the balance, continuously acquire the balance's temperature data, mass monitoring data, and tilt angle data of the x-axis and y-axis.

[0017] This application uses an electronic analytical balance as an example to calibrate data during its weighing process. Before loading an object, the electronic analytical balance is in a lever-balanced state. After loading, the torque generated by the load causes the beam to tilt. The position detector inside the balance continuously monitors the displacement of the beam or the coil connected to it, converting this displacement signal into an electrical signal. This electrical signal is then used as the error input to the control system. Combined with a PID controller, this changes the current flowing through the energized coil, causing the coil to generate a precisely controllable compensating electromagnetic force in the magnetic field of the permanent magnet, opposite to the direction of the object's gravity. This restores the beam to a lever-balanced state. The larger the current in the coil, the stronger the generated electromagnetic force. The precise mass of the object is calculated by measuring this current.

[0018] During the weighing process, even a slight tilt will change the component of gravity along the sensor's measurement axis, introducing a significant weighing error. Therefore, an inclination sensor detects the horizontal state of the balance and acquires the inclination angle data of the balance beam along the x and y axes in real time. The electromagnetic force in the electronic analytical balance directly reflects the mass of the object. During the weighing process, the current in the electromagnetic coil changes continuously until it stabilizes. An electromagnetic force balance sensor acquires the mass monitoring data of the electronic analytical balance in real time. Furthermore, temperature is also a major factor affecting weighing accuracy; therefore, a temperature sensor collects the internal temperature data of the balance in real time. In this embodiment, the acquisition frequency for inclination angle data and mass monitoring data is set to 100Hz, and the acquisition frequency for temperature data is set to 10Hz.

[0019] A weighing process is defined as a period of time during which the quality monitoring data is continuously positive. For example, if the quality monitoring data is 0 before the item to be weighed is placed, rises from 0 after the item is placed, and returns to 0 after the item is removed, then the period during which the quality monitoring data is not 0 is considered a weighing process.

[0020] Step 2: By analyzing the differences in temperature data relative to the pre-obtained optimal working temperature, the degree of temperature fluctuation, and the characteristics of temperature rise during each weighing process, a first feature value is constructed to characterize the degree of influence of temperature on the weighing results during each weighing process.

[0021] Electronic analytical balances are high-precision measuring instruments. Their electromagnetic force balance sensors, circuits, and components are easily affected by changes in ambient temperature and the heating of internal overcurrent components. Furthermore, prolonged operation can cause changes in the performance of the permanent magnet, component parameters, and materials, leading to significant reading drift. The following analysis and treatment methods address this issue.

[0022] First, operating temperature has a significant impact on component performance. Precision balances achieve the highest weighing accuracy at their optimal operating temperature, which in this embodiment is set at 20°C. When the ambient temperature fluctuates significantly or rapidly, it can cause uneven temperature changes in key components such as the permanent magnet and coil inside the balance. This leads to changes in magnetic induction intensity and the effective length of the coil, resulting in temperature drift errors. Therefore, the greater the impact on weighing accuracy, the more the actual measured temperature deviates from the optimal operating temperature or experiences rapid fluctuations during the weighing process. Taking the i-th weighing process as an example, all temperature data within the i-th weighing process and the preceding preset time (10 seconds in this embodiment) are acquired. The absolute difference between each temperature data point and the optimal operating temperature is calculated, and the average of all absolute differences is used as the temperature difference coefficient for the i-th weighing process. The larger the value, the greater the deviation of the temperature from the ideal state during the i-th weighing process. Then, the standard deviation of all temperature data during the i-th weighing process is calculated. The larger the value, the greater the fluctuation of the temperature data during the i-th weighing process, and the more unstable the weighing environment.

[0023] Furthermore, during prolonged continuous weighing, due to the limited internal space of the balance, if the accumulated heat from the coil material cannot dissipate in time, the component temperature will continuously rise, leading to a sustained increase in the local temperature of the balance. The temperature data during continuous weighing will exhibit a monotonically increasing trend, affecting the stability of the internal component performance. Therefore, the mean of all temperature data during each weighing process is calculated. The mean temperature values ​​of the i-th weighing process and the preceding N weighing processes (where N ranges from [8,15], and in this embodiment, it is 10) are arranged in chronological order and denoted as the temperature mean sequence of the i-th weighing process. The Mankendall test algorithm is used to perform trend analysis on the temperature mean sequence of the i-th weighing process, and the S-statistic output by the algorithm is denoted as the trend increase factor of the i-th weighing process. When this value is positive and larger, it indicates that the continuous heating trend in the i-th weighing process and the previous weighing processes is more obvious and more likely to affect the weighing accuracy of the balance. When this value is negative and smaller, it indicates that the cooling trend in the i-th weighing process and the previous weighing processes is more obvious and has less impact on the weighing accuracy of the balance.

[0024] In a preferred embodiment, based on the differences in temperature data relative to the pre-obtained optimal working temperature during each weighing process, the degree of temperature fluctuation, and the temperature rise trend characteristics of each weighing process and its preceding weighing processes, a first characteristic value is constructed for each weighing process. This characteristic value is used to characterize the influence of temperature data changes on the weighing results during each weighing process. The method for constructing the first characteristic value for each weighing process is as follows: obtaining the temperature difference coefficient for each weighing process; calculating the standard deviation of all temperature data during each weighing process; obtaining the trend rise factor for each weighing process; the first characteristic value for each weighing process is positively correlated with the temperature difference coefficient, the standard deviation, and the trend rise factor. This positive correlation reflects that the dependent variable increases with the increase of the independent variable and decreases with the decrease of the independent variable.

[0025] In this embodiment, the first characteristic value of the i-th weighing process is denoted as... The formula is as follows: In the formula, This is the first characteristic value of the i-th weighing process; The temperature difference coefficient for the i-th weighing process; Let be the standard deviation of all temperature data during the i-th weighing process; The trend increase factor for the i-th weighing process; For the normalization function, this embodiment uses the minimum-maximum normalization method. The minimum and maximum values ​​are determined based on the corresponding parameter set from a preset number of historical weighing processes. When the minimum and maximum values ​​are equal, it indicates that the corresponding parameters in the weighing process are relatively stable, and the likelihood of being unaffected by temperature anomalies is greater. In this case, the normalization value is set to 0. In this embodiment, the preset number is 100. If there are fewer than 100 historical weighing processes, it indicates that the electronic analytical balance has been used for a short time and the performance of each component is still good. In this case, the first characteristic value and subsequent parameters in the asymmetrical weighing process are calculated without adjusting the proportional parameters of the PLC controller.

[0026] parameter , These parameters respectively reflect the degree of deviation between the measured temperature and the optimal operating temperature during the weighing process, as well as the dynamic change characteristics of the temperature itself. The cumulative temperature rise caused by prolonged continuous operation of the balance was assessed. The results... By integrating the characteristics of temperature deviation, short-term fluctuations, and long-term trend drift, the influence of temperature state on the symmetric results was accurately quantified. The larger the value, the higher the risk of weighing error caused by temperature changes during the i-th weighing process, and the greater the impact on the weighing results.

[0027] Step 3: Based on the degree of random fluctuation of the tilt angle data of the x-axis and y-axis during each weighing process, and the degree of irregular oscillation of the quality monitoring data during decay during each weighing process, construct the second characteristic value of each weighing process. Combine the degree of synchronous change between the first characteristic value and the second characteristic value, as well as the first characteristic value of each weighing process, to determine the comprehensive abnormal characteristic value of each weighing process.

[0028] Furthermore, after detecting the tilt of the beam caused by the load, the controller of the electronic analytical balance adjusts the coil current in real time to generate a compensating electromagnetic force, causing the lever system to return to the equilibrium position. However, due to the mechanical inertia of the balance beam and the weighing pan, and the response delay in the balance adjustment between the electromagnetic force and the object's weight, the electronic analytical balance cannot achieve precise and stable tracking of the object's mass in one go. Instead, it first passes the target mass, then experiences several damped oscillations near the equilibrium position, and finally tends to a steady state.

[0029] However, the strain of elastic materials such as reeds in electromagnetic force balance sensors increases slowly over time, exhibiting creep behavior, and residual stress in metallic materials can cause thermal deformation of mechanical components. For high-precision electronic balances, creep and thermal deformation caused by the passage of time severely affect electromagnetic force balance, making it difficult for the balance beam to quickly reach equilibrium and resulting in more pronounced unstable fluctuations in the tilt angle data near the equilibrium position. Therefore, taking the tilt angle data of the x-axis during the i-th weighing process as an example, the first-order difference sequence of the x-axis tilt angle data during the i-th weighing process is calculated. Each value in the obtained first-order difference sequence reflects the amplitude of tilt angle change between adjacent sampling times. Then, the permutation entropy of the first-order difference sequence is calculated. The larger the permutation entropy, the more random the amplitude of tilt angle fluctuations of the balance beam during the process of approaching equilibrium. Similarly, for the tilt angle data of the y-axis during the i-th weighing process, the above steps can be used to obtain the corresponding permutation entropy. The average permutation entropy of the x-axis and y-axis is recorded as the tilt angle random coefficient of the i-th weighing process. The larger the value, the greater the influence of material properties on the tilt angle state of the beam during the i-th weighing process, and the worse its equilibrium stability.

[0030] As analyzed above, the balance does not measure the mass of an object precisely and stably in one go during the process of detecting the object's mass. Instead, it first exceeds the target mass, then experiences several decaying oscillations near the equilibrium position, and finally tends towards a steady state. Due to the influence of changes in the internal material properties, the more complex the dynamic balance adjustment process between electromagnetic force and gravity, the more irregular the oscillation decay characteristics of the mass monitoring data acquired by the electromagnetic force balance sensor become. Therefore, the time period in the i-th weighing process where the range of mass monitoring data within a continuous t (t is a preset monitoring duration, t is 1 in this embodiment) is less than the preset fluctuation threshold (set to 0.1% of the balance range in this embodiment) is recorded as the steady state time. The average value of all mass monitoring data after the steady state time in the i-th weighing process is recorded as the target mass. The time period from the moment the mass monitoring data first reaches the target mass to the steady state time in the i-th weighing process is recorded as the decay analysis period of the i-th weighing process. Fourier series fitting is used to fit the quality monitoring data within the attenuation analysis period, outputting the fitted curve. The extreme points in the fitted curve are obtained using derivative calculation. The absolute difference between the amplitudes of all adjacent extreme points is calculated, and the coefficient of variation between all absolute differences is recorded as the irregular attenuation coefficient of the i-th weighing process. Specifically, when the number of extreme points in the fitted curve is less than 2, the irregular attenuation coefficient of the i-th weighing process is directly set to 0, indicating that there is no irregular change in the quality monitoring data during this weighing process. The larger the obtained irregular attenuation coefficient, the more irregular the dynamic change process of the quality monitoring data during the i-th weighing process.

[0031] It should be noted that during fractional operations, when the denominator is 0, a preset minimum positive number (0.001 in this embodiment) is added to the denominator as a parameter adjustment factor to prevent the calculation from crashing due to the denominator being 0.

[0032] In a preferred embodiment, based on the random fluctuation of the x-axis and y-axis tilt angle data during each weighing process, and the irregular oscillation of the quality monitoring data during attenuation during each weighing process, a second characteristic value is constructed for each weighing process to characterize the dynamic imbalance performance of the balance during each weighing process. The method for constructing the second characteristic value for each weighing process is as follows: the permutation entropy of the first-order difference values ​​of the x-axis and y-axis tilt angle data during each weighing process is calculated; the irregular attenuation coefficient for each weighing process is obtained; the second characteristic value for each weighing process is positively correlated with both the permutation entropy and the irregular attenuation coefficient.

[0033] In this embodiment, the second characteristic value of the i-th weighing process is denoted as... Its formula is: In the formula, This is the second characteristic value of the i-th weighing process; Let be the random coefficient of the tilt angle in the i-th weighing process; Let be the irregular attenuation coefficient of the i-th weighing process; As a normalization function, this embodiment uses the minimum-maximum normalization method for normalization. The minimum and maximum values ​​are determined based on the corresponding parameter set from a preset number of historical weighing processes.

[0034] income The larger the value, the stronger the unsteady fluctuation of the tilt angle during the i-th weighing process, and the more significant the irregular attenuation of the mass monitoring data. In this case, the overall dynamic balance performance of the balance is worse, and the coordination of its internal electromagnetic force control system is worse.

[0035] Furthermore, as a high-precision instrument, the performance of internal components in an electronic analytical balance is highly susceptible to environmental changes. For example, the more pronounced the temperature drift, the greater the impact on the performance of internal components, and the more obvious the abnormal dynamic equilibrium characteristics during the weighing process. Therefore, the stronger the synchronicity between the obtained first and second characteristic values, the more significant the real-time impact of environmental temperature changes on the performance of the balance's internal components, and the greater the unreliability of the overall weighing results.

[0036] Therefore, the first and second characteristic values ​​of the i-th weighing process and the preceding N weighing processes are arranged in chronological order and denoted as the first characteristic sequence and second characteristic sequence of the i-th weighing process. The maximum mutual information coefficient between the first and second characteristic sequences of the i-th weighing process is denoted as the synchronization factor of the i-th weighing process. The larger the value, the stronger the synchronization between the temperature drift characteristic and the dynamic imbalance characteristic in the i-th weighing process, and the greater the unreliability of the overall weighing result of the balance.

[0037] As a preferred implementation, based on the degree of synchronous change between the first and second characteristic values ​​in each weighing process and the preceding weighing processes, as well as the first and second characteristic values ​​in each weighing process, a comprehensive abnormal characteristic value for each weighing process is constructed to characterize the comprehensive abnormality degree of the weighing results under multi-factor disturbances in each weighing process. The method for constructing the comprehensive abnormal characteristic value for each weighing process is as follows: the maximum mutual information coefficient between the first and second characteristic sequences of each weighing process is calculated; the comprehensive abnormal characteristic value for each weighing process is positively correlated with the maximum mutual information coefficient, the first characteristic value of each weighing process, and the second characteristic value of each weighing process. The flowchart for obtaining the comprehensive abnormal characteristic value for each weighing process is shown below. Figure 2 As shown.

[0038] In this embodiment, the comprehensive abnormal characteristic value of the i-th weighing process is denoted as... Its specific expression is: In the formula, Let be the comprehensive abnormal characteristic value of the i-th weighing process; This is the first characteristic value of the i-th weighing process; This is the second characteristic value of the i-th weighing process; Let be the synchronization factor for the i-th weighing process; As a normalization function, this embodiment uses the minimum-maximum normalization method for normalization. The minimum and maximum values ​​are determined based on the corresponding parameter set from a preset number of historical weighing processes.

[0039] income The larger the value, the stronger the overall abnormality of the weighing result in the i-th weighing process under the influence of multiple factors.

[0040] Step 4: Utilize the comprehensive abnormal characteristic values ​​to optimize the proportional term parameters of the PID controller during the next weighing process, thereby automatically calibrating the weighing results of the next weighing process.

[0041] Furthermore, conventional electronic analytical balances employ PID control for balance adjustment during the weighing process. However, their fixed, single parameters are ill-suited to adapt to complex and ever-changing working environments and real-time variations in the performance of internal components, easily leading to insufficient stability in weighing results. The obtained comprehensive anomaly value accurately assesses the comprehensive anomaly characteristics of the weighing process under multi-factor disturbances. Adaptive optimization of the PID parameters based on this value can compensate for system disturbances introduced by environmental and internal state changes in real time, significantly improving the accuracy of weighing results under non-ideal conditions.

[0042] Specifically, set the preset adjustment range for the proportional term parameter of the PID controller. ], , These are the preset minimum and maximum values ​​of the PID proportional term parameter, respectively. In this embodiment, the preset adjustment range of the PID proportional term parameter is obtained using a trial-and-error method. The larger the comprehensive abnormal characteristic value in each weighing process, the more significant the comprehensive abnormal characteristics of the weighing process under multi-factor disturbances. Therefore, a smaller proportional term parameter should be set in the next weighing process to increase stability and suppress oscillations by reducing system gain. Conversely, if the comprehensive abnormal characteristic value in each weighing process is smaller, a larger proportional term parameter should be set in the next weighing process to improve response speed and achieve rapid balance of the balance.

[0043] Calculate the proportional term parameter corresponding to the (i+1)th weighing process: In the formula, For the (i+1)th weighing process, the optimized proportional term parameter of the PID controller is used. , These are the preset minimum and maximum values ​​of the proportional parameter of the PID controller, respectively. Let be the normalized value of the comprehensive abnormal characteristic value of the i-th weighing process, wherein the normalization method is the minimum-maximum normalization method, and the minimum and maximum values ​​are determined based on the corresponding parameter set of a preset number of historical weighing processes.

[0044] During the (i+1)th weighing process, the PID controller uses the calculated proportional term parameter to adjust the balance, causing the drive coil to generate a precise compensating electromagnetic force, which restores the crossbeam to the equilibrium position, thereby achieving automatic calibration and optimization of the weighing results. Weighing using this method helps to compensate for deficiencies in weighing stability and accuracy.

[0045] Based on the same inventive concept as the above method, this application also provides an automatic calibration system for a precision balance based on multi-sensor fusion, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the above-described automatic calibration method for a precision balance based on multi-sensor fusion.

Claims

1. An automatic calibration method for a precision balance based on multi-sensor fusion, characterized in that, The method includes the following steps: During the weighing process using the balance, the temperature data, mass monitoring data, and tilt angle data of the x-axis and y-axis are continuously acquired. By analyzing the differences in temperature data relative to the pre-obtained optimal working temperature, the degree of temperature fluctuation, and the characteristics of temperature rise during each weighing process, a first feature value is constructed to characterize the degree of influence of temperature on the weighing results during each weighing process. Based on the degree of random fluctuation of the tilt angle data of the x-axis and y-axis during each weighing process, and the degree of irregular oscillation of the quality monitoring data during attenuation during each weighing process, the second characteristic value of each weighing process is constructed. Combined with the degree of synchronous change between the first characteristic value and the second characteristic value, and the first characteristic value of each weighing process, the comprehensive abnormal characteristic value of each weighing process is determined. By utilizing comprehensive abnormal characteristic values, the proportional term parameter of the PID controller is optimized for the next weighing process, thereby automatically calibrating the weighing results for the next weighing process.

2. The automatic calibration method for a precision balance based on multi-sensor fusion as described in claim 1, characterized in that, The method for constructing the first eigenvalue of each weighing process is as follows: A temperature difference coefficient is constructed based on the difference between the temperature data during each weighing process and the pre-obtained optimal working temperature. Calculate the standard deviation of all temperature data during each weighing process; Based on the temperature rise trend characteristics of each weighing process and the preceding weighing processes, a trend rise factor is constructed. The first characteristic value of each weighing process is positively correlated with the temperature difference coefficient, the standard deviation, and the trend increase factor.

3. The automatic calibration method for a precision balance based on multi-sensor fusion as described in claim 2, characterized in that, The temperature difference coefficient refers to the average absolute difference between all temperature data and the optimal working temperature during each weighing process and within the preset time period.

4. The automatic calibration method for a precision balance based on multi-sensor fusion as described in claim 2, characterized in that, The method for constructing the upward trend factor is as follows: The average temperature values ​​of each weighing process and the preceding several weighing processes are arranged in chronological order and recorded as the temperature average value sequence of each weighing process. The Mankendall test algorithm was used to perform trend analysis on the temperature mean series of each weighing process, and the output S statistic was recorded as the trend increase factor of each weighing process.

5. The automatic calibration method for a precision balance based on multi-sensor fusion as described in claim 1, characterized in that, The method for constructing the second characteristic value of each weighing process is as follows: The permutation entropy of the first difference values ​​of the inclination angle data of the x-axis and y-axis during each weighing process is calculated separately. Based on the degree of irregular oscillation of mass monitoring data during decay in each weighing process, an irregular decay coefficient is constructed. The second characteristic value of each weighing process is positively correlated with the permutation entropy and the irregular decay coefficient.

6. The automatic calibration method for a precision balance based on multi-sensor fusion as described in claim 5, characterized in that, The method for constructing the irregular attenuation coefficient is as follows: The time period in each weighing process in which the range of quality monitoring data within a continuous preset monitoring time is less than the preset fluctuation threshold for the first time is obtained, and the last moment of the obtained time period is recorded as the steady state moment. The average value of all quality monitoring data after the steady-state moment in each weighing process is recorded as the target mass. Curve fitting is performed on the mass monitoring data from the moment the target mass is first reached to the steady state moment during each weighing process, and then the extreme points in the obtained fitted curve are obtained. The coefficient of variation of the absolute difference between the amplitudes of all adjacent extreme points is denoted as the irregular attenuation coefficient of each weighing process.

7. The automatic calibration method for a precision balance based on multi-sensor fusion as described in claim 1, characterized in that, The method for constructing the comprehensive abnormal feature values ​​of each weighing process is as follows: The first and second characteristic values ​​of each weighing process and the preceding several weighing processes are arranged in chronological order and denoted as the first characteristic sequence and the second characteristic sequence of each weighing process. Calculate the maximum mutual information coefficient between the first feature sequence and the second feature sequence; The comprehensive abnormal characteristic value of each weighing process is positively correlated with the maximum mutual information coefficient, the first characteristic value of each weighing process, and the second characteristic value of each weighing process.

8. The automatic calibration method for a precision balance based on multi-sensor fusion as described in claim 1, characterized in that, The optimization formula for the proportional term parameter of the PID controller in the next weighing process is as follows: In the formula, For the (i+1)th weighing process, the optimized proportional term parameter of the PID controller is used. , These are the preset minimum and maximum values ​​of the proportional parameter of the PID controller, respectively. It is the normalized value of the comprehensive abnormal characteristic value of the i-th weighing process.

9. The automatic calibration method for a precision balance based on multi-sensor fusion as described in claim 1, characterized in that, The specific process for automatically calibrating the weighing results of the next weighing process is as follows: In the next weighing process, the PID controller uses the calculated optimized proportional term parameters to perform balance adjustment, thereby achieving automatic calibration of the weighing results.

10. An automatic calibration system for a precision balance based on multi-sensor fusion, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the precision balance automatic calibration method based on multi-sensor fusion as described in any one of claims 1-9.

Images